Technical SEO checklist for a new site

The order matters more than the list

Most technical SEO checklists are just a pile of tasks. That is unhelpful when you are launching, because some of these things gate the others. If Google cannot crawl your site, nothing downstream counts. If it crawls but cannot index, your perfect titles never get seen.

So work top to bottom. Each step assumes the one above it is done. By the end you will have confirmed — not assumed — that Google can find, read, and serve your pages.

If you want the why behind any of this, our plain-language SEO guide for founders covers the reasoning. This page is the do-it list.

1. Make sure Google can reach the right pages

Before anything clever, Google has to crawl. Crawling happens, then indexing, then serving — that is the whole pipeline, in that order, per Google's SEO starter guide.

The single most common launch disaster: a robots.txt left over from staging that says Disallow: /. That one line blocks your entire site from crawling. It happens constantly because staging environments are usually built to keep crawlers out, and nobody remembers to flip it back.

• Open yourdomain.com/robots.txt in a browser. If you see Disallow: / with nothing after it, fix it now.
• A new marketing site usually needs almost nothing in this file. Allow everything, point to your sitemap, done.
• A leftover noindex meta tag on your pages does the same damage in a different place — it lets Google crawl but tells it not to index. Check your page source for <meta name="robots" content="noindex">.

Not sure what your file should contain? Use the robots.txt generator and checker, and read the robots.txt glossary entry if the syntax is unfamiliar. The short version: robots.txt controls crawling, a noindex tag controls indexing. They are different jobs and people mix them up.

2. Serve everything over HTTPS

This is non-negotiable in 2026. If your site loads on http:// without redirecting to https://, fix it before going further.

• Confirm the padlock shows in the browser on your homepage and a few inner pages.
• Make sure http:// redirects to https:// with a 301, not a 302 and not a soft fallback.
• Pick one canonical host — either www or bare domain — and 301-redirect the other to it. Running both as live, indexable versions just splits your signals.

Most hosts (Vercel, Netlify, Cloudflare) hand you HTTPS automatically. The thing people forget is the redirect from the non-canonical host.

3. Tell Google which URL is the real one

When the same content is reachable at more than one URL, you want to name the official version. That is what a canonical URL is for: a <link rel="canonical" href="..."> tag in the page head pointing at the version you want indexed.

You need this even on a brand-new site, because duplicate URLs sneak in:

• example.com/pricing and example.com/pricing/ (trailing slash)
• URLs with tracking parameters like ?utm_source=...
• The same page reachable with and without www

There is no duplicate-content penalty — Google is clear about that. The problem is wasted crawling and Google guessing which version to show. A self-referencing canonical on each page removes the guesswork. Set the canonical to the clean, descriptive URL you actually want people to see.

While you are here: keep URLs short and readable. /blog/technical-seo-checklist beats /p?id=84792. Descriptive URLs help both users and Google.

4. Add a sitemap (optional, but do it anyway)

Sitemaps are explicitly optional — Google will find your pages through links without one. But for a new site with few inbound links, a sitemap is the fastest way to hand Google a clean list of everything you want crawled.

• Most frameworks and CMS platforms generate sitemap.xml for you. Confirm it exists at yourdomain.com/sitemap.xml and actually lists your real pages.
• Reference it from robots.txt with a Sitemap: line.
• Submit it in Google Search Console once, after you have verified ownership of the property.

Do not hand-maintain a sitemap and let it rot. A sitemap pointing at deleted or redirected URLs is worse than none.

5. Pass the mobile-first test

Google is mostly mobile-first now: the primary crawler is Googlebot Smartphone, so it judges your site by what the mobile version contains, per Google's Googlebot documentation. If content or links exist only on your desktop layout, treat them as invisible.

• Open the site on an actual phone, or use your browser's device emulation. Is the core content all there?
• Make sure Google can fetch your CSS and JavaScript — it needs the same files your visitor's browser gets. Blocking /assets/ or /_next/ in robots.txt is a classic own-goal that stops Google rendering the page properly.
• Keep your main HTML reasonable in size. Googlebot reads the first 2MB of an HTML page, so do not bury important content under megabytes of inline markup.

6. Write titles and descriptions that earn the click

The pipeline ends at serving — what shows up in results. Two quick wins per page:

• Title: unique, clear, concise. One per page, describing that specific page. Not your brand name repeated on every URL.
• Meta description: short and unique. It is not a ranking factor, and Google often writes its own snippet from your page content anyway, but a good description still influences whether someone clicks.

Skip the keywords meta tag entirely — Google ignores it and has for years. And do not obsess over word count or heading order for ranking; heading order only matters for accessibility. Write for the reader.

7. Confirm Google actually indexed you

This is the step everyone skips, and it is the only one that proves the rest worked. A few days to a couple of weeks after launch:

• Run a site:yourdomain.com search in Google. It lists the pages Google has indexed from your domain. If your key pages are missing, something upstream is broken — go back to steps 1 through 3.
• For a single page, search site:yourdomain.com/pricing to check that exact URL.
• Open Google Search Console, use the URL Inspection tool on an important page, and confirm it says the page is indexed. If not, it usually tells you why — blocked, noindex, or simply not crawled yet.

Indexing is not instant. A brand-new site with no inbound links can take a while to get noticed, since Googlebot crawls the web roughly once every few seconds on average and prioritises sites it already knows. Patience is fine. Assuming it worked without checking is not.

The five-minute pre-launch pass

If you only have time for the things that actually break sites, do these in order:

1. robots.txt does not say Disallow: /.
2. No stray noindex tag on pages you want indexed.
3. HTTPS works and http redirects to it.
4. Each important page has a self-referencing canonical.
5. After launch, site:yourdomain.com returns your pages.

The rest is polish. These five are the difference between a site that ranks and a site that is technically invisible.